Prompt 1989: What AI Tells Us About the Fall of the Berlin Wall—and What It Doesn't.

LLM services have long been more than just chatbots. For many, they’re the first port of call for “knowledge”—including knowledge about our history. However, the facts about the past aren’t quite that simple. Knowledge about the past is always selective and politically biased. This makes it all the more fascinating to look at what isn’t being told.
Let’s go back to 1989 with ChatGPT and the like. If you ask AI about the Peaceful Revolution, this is the picture you get: on one side, the SED; on the other, the civil rights movement; and in the end, the only logical goal: reunification.
Yet between the two poles familiar to our collective memory, there was a multitude of other political positions and visions. People fought for a democratic renewal of the GDR—and thus against the party dictatorship—but at the same time also against the notion that the West German system was the only possible future.
Why does this perspective surface so rarely today? And what does that reveal about the data from which AI learns?
Taking as its starting point ROBOPSY, the media art installation by Margarete Jahrmann that examines historical reconstructions through language models, artistic research meets contemporary history and an analysis of the present.
Together with civil rights activist Hans Misselwitz and journalist Nine Christine Müller, we ask how memory is formed in the age of generative AI. And why it could be crucial for our democracy that we don’t just pass on the most likely story or stories.

What do you think? I tried to write this in a very pop style (because I thought that would reach the target audience). Is that right? Would you do it differently? I’m counting on your experience here.